A Hedonic Analysis Of The Value Of Urban Housing Based On Spatial Durbin Model

As a kind of heterogeneity goods, foreign scholars widely used Hedonic Price Model (HPM) to explore the formation mechanism of housing price. The exsiting research achievements show that, even if considering the variables of location, HPM can result in biased results. That’s because HPM ignores the spatial effect of housing price and its characteristics. Therefore, based on spatial econometrics, this paper tries to use the recently developed Spatial Durbin Model (SDM) to study of the hedonic price of urban housing, and reveal the space effect of hedonic price.First of all, based on the literature review of the study on the research progress, theory and methods of hedonic price, this paper points out that SDM is suitable for the study of hedonic price of housing. According to the fact that SDM haven’t been used for the study of hedonic price in China, this paper tries to conduct exploratory research. Then, this paper takes Hangzhou, Zhejiang Province as an example. The Exploratory Spatial Data Analysis (ESDA) and Universal Kriging Interpolation is applied to reveal the spatial variation of the hedonic price based on face-to-face survey. And then, the model is estimated by HPM、Spatial Lag Model (SLM)、Spatial Error Model (SEM) and SDM in order. Finally, through the selection of models, SDM is accepted to analysis the space effect of hedonic price as well as its reason.Based on it, this paper concludes that:(1) The feature of spatial distribution of urban housing in Hangzhou is changing from ’single-core’ to ’multi-core’. The housing price declines gradually from CBD to the suburban areas, meanwhile with several exceptions affected by the emerging plate, such as Qiaoxi Plate、Shenhua Plate、Dingqiao Plate and so on;(2) The Global Moran’I is0.34, and significant at the1%level, which shows that spatial autocorrelation exists on housing price of Hangzhou. That means the price of one house has a positive effect on another one nearby;(3) SDM clearly reveals the influence on housing price with different characteristics and quantify these overflow price, and it also shows higher degree of model fit and better explanation than HPM and other two spatial econometrics model(SLM&SEM);(4) The result of SDM shows, six spatial lag variables are selected into the model, including construction area, decoration, garage, surrounding, educational facilities and distance to Wulin Square, which have spatial diffusion among housing prices.